gpu_type: nvidia-tesla-a100
gpu_num: "8"
gpu_memory: "80"  # "80"
# resource should scale accordingly to the gpu_num requested, cpu_num is recommended to be gpu_num x 11, same applies to the memory
cpu_num: "88"
memory: "880.0"

# Add required resource and ENV vars for EFA (note: Only works on AWS, and only used for distributed training)
efa: False
# efa_count: "4"
# huge_pages: "21122Mi"

priority: gpu-high

pytorchjob:
  replicas: 1 # number of replicas to run

custom_node_labels:
  snap.com/zone: us-west-2c
  snap.com/ami_type: al2023

volcano_queue: omni-id

# custom_node_labels:
#   # use_case: research 
#   snap.com/zone: us-west-2c


script:
#  pre_run_event: "cd /nfs/yli16/effvideo && pip install -r requirements.txt && pip uninstall -y apex && wandb login --relogin --host=https://snap.wandb.io local-5c7725e9c8a991d399dfee329e7c5c9f7b33de8e"
  #pre_run_event: "cd /nfs/ywang29/video_reward/vila_ft && wandb login --relogin --host=https://api.wandb.ai 24d6afefd46915ea938dca1af6caf95ae0288a19 && bash ./environment_setup.sh lift && source activate && conda activate lift && pip uninstall lift && pip install huggingface_hub==0.31.1 && pip install accelerate==0.32.0 && pip install peft==0.10.0"
  pre_run_event: "cd /nfs/ywang29/SnapVideo &&  /nfs/ywang29/Reward_finetuning/SAGE-GRPO/pre_run_2_10.sh && wandb login --relogin --host=https://api.wandb.ai 24d6afefd46915ea938dca1af6caf95ae0288a19"

  #pre_run_event: "bash ./environment_setup.sh lift && conda activate lift && pip uninstall lift && pip install huggingface_hub==0.31.1 "
#  pre_run_event: "cd /nfs/yli16/snapgenv-v2-i2v && pip install -r requirements.txt && pip uninstall -y apex diffusers coremltools && pip install coremltools && wandb login --relogin --host=https://snap.wandb.io local-5c7725e9c8a991d399dfee329e7c5c9f7b33de8e"
#  pre_run_event: "cd /nfs/yli16/snapgenv-v3 && pip install --upgrade pip && pip uninstall -y diffusers coremltools peft && pip install coremltools peft && wandb login --relogin --host=https://snap.wandb.io local-5c7725e9c8a991d399dfee329e7c5c9f7b33de8e"
  # command: "cd /nfs/ywang29/Reward_finetuning/VideoX-Fun && source activate && source /nfs/ywang29/Reward_finetuning/VideoX-Fun/scripts/wan2.1/env.sh  &&  /nfs/ywang29/Reward_finetuning/VideoX-Fun/validation_full_vbench2.0_control.sh  && while :; do sleep 1000; done"
  command: "cd /nfs/ywang29/Reward_finetuning/SAGE-GRPO/ && MAX_PROMPTS_PER_DIM=0 SAMPLES_PER_PROMPT=auto bash run_sage_vbench2_generate.sh yw004-flow-baseline 100 && RUN_SETUP=true  bash run_vbench2_eval_only.sh yw004-flow-baseline 100"


  jobs:
    - name: ywang29-vbench2-eval-1

project_name: mobile_video
project_support_alias: ywang29
team: creative_vision
cost_team: generative_ai
cost_feature: generative_ai_development
cost_sub_feature: video_generation

#docker_image: 440036398022.dkr.ecr.us-west-2.amazonaws.com/facecraft-ml:efficient_video_ywu6-6b32aee
#docker_image: 440036398022.dkr.ecr.us-west-2.amazonaws.com/facecraft-ml:efficient_video_ywu6-3942a73
# docker_image: 440036398022.dkr.ecr.us-west-2.amazonaws.com/facecraft-ml:efficient_video_ywu6-eb1a4a7
# docker_image: 763104351884.dkr.ecr.us-west-2.amazonaws.com/pytorch-training:2.7.1-gpu-py312-cu128-ubuntu22.04-ec2
docker_image: 763104351884.dkr.ecr.us-west-2.amazonaws.com/pytorch-training:2.9.0-gpu-py312-cu130-ubuntu22.04-ec2

mount:
  - type: s3
    name: snap-research-cv-code
    #prefix: logs/ # optional bucket's prefix
    #region: us-west-2 # optional bucket's region (by default cluster's region will be used)
    mount_path: /s3-code
  - type: fsx
    name: "fs-0aec9ce19ecc66a45"
    mount_path: /nfs
  - type: fsx
    name: fs-0a730854e147a7bc2  # generative-one-shot
    mount_path: /nfs-generative-one-shot
  - type: fsx
    name: fs-0edcd5941d62ce319
    mount_path: /snaphub
  - type: s3
    name: snap-genvid
    mount_path: /snap-genvid
  - type: s3
    name: snap-genvid-us-east-2
    mount_path: /snap-genvid-us-east-2
    

#mkdir: /nfs/ywu6/logs